Three convergent lines of research will be continued. (1) Application of event history analysis and time series analysis to the examination of the natural history of addiction, wherein significant events in the initiation, maintenance, and interruption of addiction will be assessed, with attention to treatment and criminal justice system intervention effects. (2) Extension of current work on the development of drug use prevalence estimation modeling techniques using continuous-time Markov models and time series techniques to narcotics addict and cocaine abuser career data. Eventual goals are to generalize the application of these techniques to larger populations as well as to populations using other types of drug. (3) Evaluation of the reliability and validity of addict self-report and identification of factors associated with differential degree of reliability and validity. These three lines of research will initially be conducted on narcotics addicts and subsequently will be extended and replicated on cocaine user populations in the later two years of the proposed award period.